Can Elicitation Methods Increase the Precision of Fair Value Estimates?

This dissertation is motivated by recent changes in financial reporting regulation effected by the adoption of IFRS in Europe, Australia, and Canada, and SFAS 142 (FASB 2001) and SFAS 157 (FASB 2007 and 2011) in the U.S., that significantly increased users’ exposure to fair values. The implementation of the fair value hierarchy, as well as the switch from amortization to impairment testing of goodwill, highlighted problems with auditing highly complex, judgment-dependent and inherently uncertain fair values. There is a concern that such fair values may not always be auditable, and that requiring auditors to provide positive assurance on them may necessitate changes to the financial reporting model. The dissertation consists of two parts, the process study and the elicitation study. The process study, using the audit of goodwill/cash generating unit (CGU) impairment under IFRS as a specific example, provides quasi-experimental evidence about the fair value auditing process which can help to better understand and improve the auditing of complex fair values. The study relies on an analysis of verbal protocols to develop an understanding of how auditors and valuation specialists deal with the task. The study finds that for all of the participants who developed an auditor’s range, the width of the range is many times the audit materiality, and intervals for the experienced auditors are narrower on average than those for junior auditors. There are signs of possible issues with both interpretation and application of fair value auditing and accounting standards across all groups of the participants. At least some of the issues with application of the standards appear to be related to judgmental shortcuts (heuristics) which have not been researched in a valuation task context in prior auditing literature. Some of the experienced and junior auditors do not appear to have a complete grasp of the applicable valuation methodology. Finally, the results shed light on the division of responsibilities between assurance and valuation groups and the use of third party experts when auditing fair value impairments. The process study contributes to the literature by obtaining direct quasi-experimental evidence on auditors’ and valuation specialists’ process when they perform a fair value auditing task, and investigating the process differences among auditors with different levels of experience and experts. The objective of the elicitation study is to develop techniques that can be used by auditors and valuation specialists when auditing complex fair values, by experimentally testing elicitation methods for fair value models’ parameters. The study tests two probability distribution elicitation methods - the cumulative distribution function (CDF) method and the credible interval (CI) method. Quantitative analysis performed in this study indicates that the CDF method has a potential to improve the participants’ unaided judgment regarding fair value intervals, at least for junior auditors, while the CI method does not yield similar improvement. When the two methods are compared to each other, the CDF method proves to be more effective for experienced and junior auditors, while the opposite is true for valuation specialists. The distributions developed with the help of the CDF method are subjected to the effects of anchoring heuristic to a lesser degree than those built using the CI method. Qualitative analysis based on verbal protocols in the elicitation study indicates that the CDF distribution elicitation method surpasses the CI method for the purposes of quantification of uncertainty inherent in complex fair value estimates. The study contributes to the literature by combining auditing and elicitation research in fair value auditing settings, and has a potential to improve the practice of auditing of goodwill and possibly other complex fair values, by providing information for the development of relevant decision aids